Organic photochemistry with 6.7-eV photons: .gamma.,.delta

Organic photochemistry with 6.7-eV photons: .gamma.,.delta.-unsaturated ketones. William J. Leigh, and R. Srinivasan. J. Am. Chem. Soc. , 1982, 104 (1...
0 downloads 0 Views 760KB Size
4424

J. Am. Chem. SOC.1982, 104, 4424-4429

Organic Photochemistry with 6.7-eV Photons: y,&Unsaturated Ketones William J. Leigh and R. Srinivasan* Contribution from the IBM Thomas J. Watson Research Center, Yorktown Heights, New York 10598. Received November 19, 1981

Abstract: The photochemistry of a series of acyclic, aliphatic y,&unsaturated ketones in pentane solution with 185-nm light has been investigated. The variety of products formed can be rationalized as arising from discrete reactions of the individually excited chromophores. Neither intramolecular energy transfer nor internal conversion leading to population of the lowest (n,r*) singlet state are competitive with product formation at 185 nm. The mechanism of alkyne formation in the solution-phase photolysis of these and terminal alkenes at 185 nm is elucidated by using a deuterated derivative, and the reaction is postulated to proceed from the olefinic r,r* singlet state.

The photochemistry of bichromophoric organic molecules in the far ultraviolet'-* displays distinctly different characteristics from what has generally been observed at longer wavelength^.^ These characteristics result from (i) the reactive chromophore (at 185 nm) being transparent a t longer wavelengths and reacting faster than internal energy transfer can take place', and (ii) photochemistry ensuing from upper excited states which are the result of through-bond and through-space mixing of the isolated chromophores.2 A further result of ii is that the overall photochemistry cannot be accounted for in terms of the sum of the separate parts. Published work in this area encompasses olefin/olefin'- and olefin/cyclopropyPtype systems. In this report we examine the far ultraviolet photochemistry of linear, nonconjugated ketone/olefin bichromophoric systems, la-d.

1

a, R, = CH,, R, = R, = H; b, R, = R, = R, = H; c, R, = R, = D,R, = H; d, R, = R, = H, R, = CH,

The long wavelength photochemistry of y,d-unsaturated ketones has been extensively studied4 and can be simply characterized as that of an n,r* aliphatic ketone in the presence of an alkene (Scheme I). The (initially excited) n,a* singlet state of the carbonyl moiety undergoes intramolecular oxetane formation, in competition with intersystem crossing to the n,r* triplet state followed by intramolecular energy transfer to the olefinic end of the molecule and resultant cis-trans isomerization. The formation of 2 and 3 as well as ca. 67% of the total observed cis-trans isomerization is presumed to involve the intermediacy of an intramolecular exciplex.3as" The excited-state manifold in this system, depicted in the energy level diagram in Figure 1, is expected to be complex in the 185-nm region. The ultraviolet absorption spectra of both aliphatic ketones and alkenes in hydrocarbon solution reveal absorption bands in this r e g i ~ n . ~In aliphatic ketones, this band corresponds to excitation to the second excited singlet state and has been identified (1) Srinivasan, R.; Brown, K. H.; Ors, J. A.; White, L. S.; Adam, W. J . Am. Chem. SOC.1979, 101, 7424. (2) (a) Srinivasan, R.; Ors,J. A.; Brown, K. H.; Baum, T.; White, L. S.; Rossi, A. R. J. Am. Chem. SOC.1980,102,5297. (b) Srinivasan, R.; Baum, T.; Brown, K. H.; Ors,J. A.; White, L. S.;Rossi, A. R.; Epling, G. A.; J. Chem. SOC.D 1981, 913. (3) (a) Morrison, H. Acc. Chem. Res. 1979, 12, 383. (b) Morrison, H. Org. Phorochem. 1979, 4, 143. (4) (a) Srinivasan, R. J . Am. Chem. SOC.1960, 82, 775. (b) Morrison, H. Tetrahedron Lett. 1964, 3653. (c) Morrison, H. J . Am. Chem. SOC.1965, 87, 932. (d) Yang, N. C.; Nussim, M.; Coulson, D. R. Tetrahedron Lett. 1965,1525. (e) Kurowsky, S. R.; Morrison, H. J. Am. Chem. SOC.1972,94,

507. ( 5 ) 'DMS UV Atlas of Organic Compounds"; Plenum Press: New York, 1968; Vol. 4.

Scheme I

0.'

I

1ISC

+

,x/..AR

/

2

3

cis-timi

llomnlrrtlon

as the no,3s(R) Rydberg or (more likely) n,u* valence shell transition.6 The exact identity of this state is subject to controversy; at any rate, the solution-phase photochemistry of aliphatic ketones in the far-UV has not been investigated in detail. In monoand disubstituted aliphatic olefins, the absorption in this region is composed of two bands, corresponding to excitation to the r,3s(R) Rydberg and r,r* singlet states; the relaxed (orthogonal) r,r*state is believed to be lowest in energy. Fluorescence emission studies suggest that internal conversion from the r,3s(R) state to the r,r* singlet state is extremely rapid in mono- and disubstituted systems.' While the photochemistry of tri- and tetrasubstituted acyclic alkenes has been extensively investigated,* that of simple terminal alkenes has received limited attention. Given the complexity of the excited state manifold in this system, there are several possibilities that may occur upon excitation of 1 with 185-nm light. Photobehavior may be dominated by carbonyl photochemistry, which may result in either unique reactivity or enhanced n,r* reactivity depending on how fast internal conversion to the lowest (n,**) singlet state occurs. On the other hand, olefin photochemistry may dominate, providing that intramolecular energy transfer to the carbonyl group (leading again to enhanced n,r* photoreactivity) is not competitive. Also, the possibility exists that this spectral region contains transition(s) to mixed states, in which case unique photochemistry may again be observed. Our results enable some distinction to be made between the various possibilities outlined above, as well as a definition of the photochemical behavior of this system and simple alkanes at 185 nm.

Results The ultraviolet absorption spectra in the region below 220 nm of trans-la and l b in pentane solution are shown in Figure 2. The photochemistry of trans-5-hepten-2-one ( l a ) in pentane solution at 185 (+254) nm is summarized in eq 1. The oxetanes 2a and 3a which were also formed and a fraction of the cis-la obtained were established to result from irradiation with the (6) Robin, M. B. "Higher Excited State of Polyatomic Molecules"; Academic Press: New York, 1975; Vol. 2. (7) Hirayama, F.; Lipsky, S.J . Chem. Phys. 1975, 62, 576. (8) Kropp, P. J. Org. Photochem. 1979, 4, 1.

0002-1863/82/ 1504-4424$01.25/0 0 1982 American Chemical Society

J. Am. Chem. Soc., Vol. 104, No. 16, 1982 4425

Organic Photochemistry with 6.7-eV Photons

5a

4

x

62""

1

6 2 "h

1 "

2%

13%

(1) 254-nm component of the lamp output. No other product besides 2a, f, and cis-la was detectable in the 254-nm photolysate. cis-la is apparently comparatively unreactive a t 185 nm (as well as at 254 nm"), since concentration vs. time plots for the formation of 2a and 4-7 are noticeably curved. The yield of cis-la was calculated (by GC) after 15-min photolysis, while those of 4-7 were calculated from the initial slopes of the concentration vs. time plots in Figure 3, relative to the slope of the [trans-la]vs. time plot after correcting for the 254-nm component. The nonvolatile products were shown by G C to consist of at least 15 high-boiling materials, none in appreciable yield; the total yield was determined after 120-min photolysis. Photolysis of 5-hexen-2-one ( l b ) at 185 (+254) nm in pentane solution gave the product mixture summarized in eq 2. Non-

t

8o

lb

Figure 1. Excited-state energies for 5-hexen-2-one ( l b ) .

volatile products (40%), determined at 19% conversion, were not isolated but were shown (by GC) to consist of more than 15 products, none in greater than 2% yield. The oxetanes were shown to result entirely from irradiation with the 254-nm component of the lamp output. No other products besides 2b and 3b were detected in a 254-nm photolysis under identical conditions. Straight lines were obtained in all cases from concentration vs. time plots for the products and starting material, and yields were determined from the calculated slopes, after correcting for the 254-nm component. Other minor products included 1-butene (2%) and ketene (trapped with methanol, ca. 2%). Photolysis of l b at 185 (+254) nm as a film at -70 OC under high vacuum was camed out to determine the formation of gaseous products. Fractional distillation of the trapped materials lead to the identification (by IR) of acetaldehyde, acetylene, ketene, 1,3-butadiene, and 1-butene as minor products. The bulk of the photolysate was similar in all respects to that obtained in the solution phase photolysis. Photolysis of 5-hexen-2-one-6,6-d2 ( I C ) at 185 (+254) nm in pentane solution gave a product mixture which was identical with that obtained from the photolysis of l b (see eq 3). Gas chrok

2

+%--

2-

. 1-

180

190

2-

.

IC

16 7 0%

3c IO".

10

9C

2096

5 "0

I C

10~0

(3) 180

190

matographic isolation of 6,followed by infrared and mass spectral analysis, showed it to be ca. 10% deuterated in the terminal position?JO Recovered 5-hexen-2-one was examined by deuterium

Figure 2. Ultraviolet absorption spectra (pentane solution) of trans-la and l b .

(9) Comparison of the mass spectra of 6 isolated from the photolyses of l b and l e indicates that the latter is less than about 10%deuterated. For a review of the mass spectra of alkynes, see: Dolejsek, Z.; Hanus, V.; Vokac, K. Adu. Mass Spectrom. 1966, 3, 503-514. (10) The infrared spectrum clearly shows two bands attributable to alkyne C-H and C-D stretches at 3295 and 2600 an-',respectively, with an intensity ratio of 13295/Im = 12.0. It has been determined that the ratio of intensities of the C-H and C-D stretches in 3,3-dimethyl-l-propyne and 3,3-dimethyl1-propyne-1-d is 0.9;" this leads to an estimate of ca. 7% deuterium at the 6-posrtion In 6. (1 1) (a) Hoffmann, V.; Stehlik, G.; Zed, W. Z. Nafurjorsch.,A 1970,25A, 572. (b) Gastilovich; Shigorin; Komarov O p f . Spektrosk. 1965, 19, 354.

N M R spectroscopy. The proton-decoupled spectrum consisted of three singlets at 6 2.28,4.99, and 5.85 in the ratio 1:8:1. These signals correspond presumably to deuterium in the 4-, 6 - , and 5-positions, respectively, and thus indicate that the recovered starting material (after 40% apparent conversion) consisted of IC (70%), 5c (lo%), and 10 (20%). No detectable scrambling of deuterium was evident in I C recovered after a 254-nm photolysis run under comparable conditions. Product yields, not corrected for 254-nm products, were determined by G C at 40% conversion, taking deuterium scrambling (leading to 5c and 10) into account.

4426 J. Am. Chem. SOC.,Vol. 104, No. 16, 1982

Leigh and Srinivasan

Table 1. Rates of Akenone Disappearance and Product Formation in the Photolyses of 1 and 1-Octene" kZb

kdisb

185 (+254)e

254d la lb Id 1-octenef

4.83 3.00 6.00

26.7 8.50 19.33 1.20

254d

185 (+254)e

1.17 0.72 3.20

2.18 1.05 5.58

k4blC

1.63

kdbiC

k,bic

k,blC

1.10

0.32 0.72

0.32 0.25

kBbOc

kgbpc

1.05 0.82

0.62

" 0.05 M, degassed pentane solution, 15 "C, 40-Wlamp.

xlOS s; error is +5%. kl-octsne = 1.28.

Obtained from the slopes (or initial slopes for la) of concentrations vs. time plots. These products are formed only at 185 nm. In nm. Vycor filter, %T,,, = 69. e In nm. Unfiitered. oxetanes were shown to result primarily from irradiation with the 254-nm component of the lamp output. No other products besides 2d and 3d were detected in a 254-nm photolysis under identical conditions. The yields of 7 and 8d were determined by G C relative to reacted starting material at 12% conversion, while the yield of nonvolatile products (not isolated) was determined at 25% conversion (corrected for 254-nm component). Straight lines were obtained from concentration vs. time plots for all products and starting material up to 20% conversion. Rate data for disappearance of the starting material and formation of the products are collected for la,b,d in Table I for both 185 (+254)-nm and 254-nm photolyses. The solution concentrations (0.05 M) were such that all of the incident radiation at 185 nm was absorbed, the rates of formation of the products should therefore be proportional to their quantum yields. A quantum yield of 0.015 was determined for the formation of 2a from trans-la.& The lamp output at 254 nm is estimated to be ca. 6 times greater than its output at 185 nm. Photolysis of 1-octene at 185 nm in pentane solution lead to the formation of four products (eq 5). Three of these, formed

3 [t-la]

(arb. unii 3

2

-

2

185 nm

pentane*

I

I

50

100

150

B.

*b. UIiits)

54 ?o/,

I

-

t (min.) Figure 3. Concentration versus time plots for the photolysis (185 (+254) nm) of trans-la (pentane solution, 0.05 M, 15 "C): A, disappearance of truns-la; B, formation of 2a, 4, Sa, and 6.

8c and 9c were not isolated in this case.

Photolysis of 5-methyl-5-hexen-2-one (la) at 185 (+254) nm in pentane solution gave the product mixture shown in eq 4. The

Id

+.%E+

+

L*Me + + ; ; ',p": 7

8d

3 06

9 "h

(5)

-*H 18%

in yields of 32%, 16%, and 6%, were determined by G C / M S to be isomeric with the starting material. The fourth product (18%) was identified as 1-octyne on the basis of a comparison of its IR and mass spectra and G C retention times with those of an authentic sample. Rate data for the photolysis of 1-octene are collected as well in Table I.

Discussion As Figure 2 reveals, the ultraviolet absorption spectra of trans-5-hepten-2-one (la) and 5-hexen-2-one (lb) in the region below 220 nm are featureless and offer no indication of the presence of mixed transitions; they are probably best viewed as the superposition of the So S2 and So SI absorption spectra of the isolated carbonyl and olefinic chromophores. Since absorption in this region is about a factor of 10 greater in olefins (ela5 -lo4) than in alkyl ketones (elas N ~ O ~one ) , can ~ conclude that the bulk of the incident radiation at 185 nm is absorbed by the olefinic portion of the molecule. As expected, the spectrum of trans-la extends to longer wavelengths than does that of lb, in accord with the known effect of increasing alkyl substitution on the absorption spectra of aliphatic olefins.6 The photochemistry of the four compounds (la-d) investigated at 185 nm is generally analogous to that of acyclic l,5-dienesl2 and can be rationalized as resulting primarily from excitation of the isolated carbon-carbon double bond. Thus, cis-trans isomerization (accounting for 62% of the isolable products in the case of trans-la) provides the dominant mode of excited-state decay, while [ 1,3]-hydrogen migration (leading to 4 from l a and 8 from lb-d), [ 1,2]-hydrogen migration, and [ 1,3]-acylmethyl migration are of secondary importance. In contrast to our proposed [ 1,2]-H shift mechanism to account for the formation of 9 and 10 from la,c via the carbene intermediate shown in eq 6, it is noteworthy

-

+

+

I

t (min.)

.10

C P , , i~ome's ( 3 )

-

2 7 "0

(4)

(12) Manning, T. D.

R.; Kropp, P. J. J . Am. Chem. SOC.1981,103, 889.

J. Am. Chem. Soc., Vol. 104, No. 16, 1982 4421

Organic Photochemistry with 6.7-eV Photons Scheme I1

0

that analogous allylcyclopropane formation in the photolysis of 1,5-dienes has been shown to result from initial [1,2]-allyl migration followed by closure of the resulting 1,3 biradical.‘* However, [1,2]-H(D) migration is the only pathway which accounts for deuterium scrambling (into the 5-position) in the photolysis of lc. Although the corresponding methylcyclopropane was not detected in the photolysis of either l a or Id, one cannot rigorously exclude the possibility of its presence in minor amount. [ 1,3]-Acylmethyl migration (leading to 5 from la,c) has been observed previously upon short wavelength irradiation of a cyclic y,b-unsaturated ketone13 and the conclusion that this reaction proceeds from an upper excited state is substantiated by our results. The photoelimination of R H (R = CH3 or H), leading to alkynones 6 and 7,is noteworthy in that it has not been previously recognized as a common photoreaction of alkenes in solution, though well-known in the gas phase.14 The identification of 1-octyne as a major product (18%) in the 185-nm photolysis of 1-octene clearly indicates its generality, and the Occurrence of this reaction in the photolyses of l a - d is thus not due to any special interaction between the olefinic and carbonyl groups. There are at least two possible mechanisms for the formation of 6,from 5-hexen-2-one (lb), and these are illustrated in Scheme 11. Path “a” involves elimination of H2 (or RH) across the double bond, leading to the alkynone directly. This is the only mechanism by which 7 can be formed in the photolysis of trans-la. Path “b” involves loss of RH from the 6-position (shown formally as leading to a vinylidene intermediate15) followed by, or in concert with, [1,2]-hydrogen migration from the 5- to the 6-position. Both pathways are known to occur in the gas-phase photolyses of olefin^,^^^'^ although considerably higher energy excitation is usually necessary. The present study indicates that both occur as well in solution, though with varying importance depending upon the degree and pattern of substitution at the double bond. Mass9 and infraredl0 spectral analysis of the 5-hexyn-2-one (6) isolated from the photolysis of 5-hexen-2-one-6,6-d2 (lc) and the absence of 6 in detectable amount in the photolysis of 5methyl-5-hexen-2-one (Id) indicates that elimination proceeds at least 90% via path “b” in terminal (mono- and 1,l-disubstituted) olefins, while the formation of both 6 and 7 in the photolysis of 5-hepten-2-one ( l a ) indicates that alkyne formation proceeds at least 50% via the alternate pathway (“a”) in the case of 1,2-disubstituted olefins. It is important to recognize that the formation of 7 from Id requires [1,2] migration of a methyl group.17 Alternate scission of the C& bond, leading to acetylene and 2-butanone from l b presumably also occurs, although to a lesser extent. Acetylene was detected as a minor product in the photolysis of l b as a frozen film. The nonvolatile material formed in significant yields in the photolyses of l a - d could have arisen from a number of sources, though no single compound was formed in sufficient yield to enable isolation. The formation of ketene/ 1-butene and acetaldehyde/ 1,3-butadiene as minor products in the photolysis of l b suggests that a cleavage is a minor reaction pathway, arising from population of the weakly absorbing carbonyl n,u* state, so it may be that the nonvolatile materials are formed as a result of the production of free radicals.

L

H

A

The differences in the rates of oxabicyclohexane (2) formation in the unfiltered and Vycor-filtered photolyses (Table 1) can be accounted for by the reduction in intensity of the 254-nm light as a result of the transmittance characteristics of the Vycor filter. The implication that very little or none of these products (Le., 2 and 3) are formed at 185 nm18 leads to the important conclusion that the lowest (n,a*) state of 1 is not populated to any significant extent upon absorption of 185-nm light. This means, first, that intramolecular energy transfer from the olefinic S1state to the carbonyl n,a* state must be slow relative to the rate of reaction of olefinic singlets, in spite of the facts that such a process would be exothermic by at least 30 kcal/mol and that a significant fraction of molecules must be in such an orientation when excited to enable energy transfer to occur at a rate considerably faster than the normal bimolecular diffusion-controlled rate (ca. 2 X 1OloM-’ s-I in n-pentane19). However, while the lower limit for the rate constantsZofor formation of the olefin-derived products from 1 is only 1Olos-l, the poor overlap between alkene emission7 and ketone absorption spectra may slow the rate of intramolecular r,a* n,r* singlet-singlet energy transfer considerably due to Franck-Condon f a ~ t 0 r s . lSecond, ~ the apparent absence of n,a* products upon 185-nm excitation may imply that carbonyl S2+ S1internal conversion is slow as well in these compounds, although caution should be exercised in this interpretation. Despite the fact that the large energy difference (30 kcal/mol) between the zero-zero bands of S2 and S1may make rapid internal conversion ~ n l i k e l y , only ’ ~ 10% of the light at 185 nm is absorbed by the carbonyl moiety, and the yield of n,a* products formed as a result of S2 SI internal conversion would be within our limits of error.’’ The observation of a-cleavage products (ca. 2%) in the 185-nm photolysis of l b does indicate, however, that S2 SI internal conversion is, at most, competitive with reaction from the carbonyl n,u* state. Finally, there remains the question as to which excited state@) is involved in the olefin-derived photoreactions observed for la-d. It has been suggested that [1,3]-hydrogen migration and [1,2] migrations leading to carbenes are derived from the Rydberg state.’ If this is the case, then it may be that the alkynes are derived from the a,a*state. This is consistent with the assigned ordering of the a,3s(R) and *,a*states in alkyl olefins.6 Heavily alkylated olefins, in which the lowest singlet state is r,3s(R), do not give alkynes upon photolysis but instead yield [ 1,3]-hydrogen migration and carbene-derived products and react with nucleophiles. Upon proceeding to tri-, di-, and monosubstituted olefins, the a,3s(R) state increases in energy relative to the a,r*state, and these reactions decrease in importance (especally nucleophilic trapping) at the expense of alkyne formation.

-

-

-

Experimental Section Infrared spectra were recorded in carbon tetrachloride solution on a Beckman Acculab 6 spectrometer and are reported in wavenumbers. ‘H N M R spectra were recorded on a Varian T60 spectrometer and are reported in parts per million downfield from Me&. Inverse gate proton-decoupled *H N M R spectra were recorded on an IBM Instruments ~~~

(13) Crandall, J. K.; Mayer, C. F.; Arrington, J. P.; Watkins, R. J. J. Org. Chem. 1974, 39, 248. (14) McNesby, J. R.; Okabe, H. Adv. Photochem. 1964, 3, 157. (15) Krishnan, R.; Frisch, M. J.; Pople, J. A,; Schleyer, P. von R. Chem. Phys. Left. 1981, 79, 408. (16) Okabe, H.; McNesby, J. R. J . Chem. Phys. 1962, 36, 601. (17) (a) Gilbert, J. C.; Luo, T.J . Org. Chem. 1981, 46, 5237. (b) Wolinsky, J.; Clark, G. w.; Thorstenson, P. C. Ibid. 1976, 42, 745.

0

~

(18) 2b and 3b are formed in ca. 2% chemical yield upon photolysis of a degassed, 0.05 M pentane solution of 5-hexen-2-one (lb) with a defocused argon fluoride laser (A = 193 nm; fluence = ca. 15 mJ/cm2). The product distribution is otherwise identical with that obtained in the 185 (+254) nm

irradiations: Leigh, W. J.; Srinivasan, R. unpublished results. (1 9) Turro, N. J. ‘Modern Molecular Photochemistry”, Benjamin/Cummings Publishing Co.: Menlo Park, CA, 1978; Chapter 9. (20) Estimated from the absorption spectrum of 2-per1tene.~

4428 J. Am. Chem. SOC.,Vol. 104, No. 16, 1982

Leigh and Sriniuasan

NR80 spectrometer, calibrated with a benzene-d, external reference, and are reported in parts per million downfield from Me,Si. Ultraviolet absorption spectra of GC purified materials were recorded in pentane (Baker Photrex-grade) solution on a Cary 17D spectrometer and are reported in nanometers. Mass spectra were recorded on a HewlettPackard 5995A gas chromatography/mass spectrometer equipped with either 3% Carbowax 20M (12 ft X 0.25 in.) or 3% OV-101 (12 ft x 0.25 in.) glass columns. Analytical and preparative gas chromatographic separations were carried out on a Hewlett-Packard 5750B gas chromatograph equipped with a thermal conductivity detector and the following columns: (a) 10% Carbowax 20M, 14 ft X 0.25 in.; (b) 10% OV-101, 14 ft X 0.25 in.; (c) 20% Apiezon L, 14 ft X 0.25 in. t"4-Hepten-2-one ( l a ) was prepared by the method of Kimel and Cope.21 The product was distilled through a 12411. spinning band column, affording material (bp 142 "C) which was >99% pure by GC and contained less than 5% of the cis isomer. 5-Hexen-2-one (lb) was Eastman reagent grade and was purified by passage through an alumina column followed by two fractional distillations over anhydrous sodium sulfate through a 15-in. Vigreux column. The material (bp 121 "C) was >99% pure by GC. 5-Hexen-2-one-6,6-d2 (IC) was prepared by the Wittig reaction of (3-(2-methyl-1,3-dioxan-2-yl)propyl)triphenylphosphoniumbromidez2 and paraformaldehyde-d2. The phosphonium salt (13.7 g, 0.03 mol) and anhydrous ether (125 mL) were placed in a 250" round-bottom flask fitted with condenser, magnetic stirrer, and nitrogen line. A 2.4 M solution of n-butyllithium (17 mL, 0.035 mol) was added, and the mixture was stirred for 2 h at room temperature. Paraformaldehyde-dz (CPL; 1.8 g, 0.056 mol) was added to the orange mixture, and stirring was continued for a further 24 h. The creamy, slightly yellow mixture was filtered through Celite, the filter cake was washed repetitively with ether, and the solvent was evaporated to yield a yellow oil (2.2 g). The oil was redissolved in ether (10 mL), 2 N hydrochloric acid (10 mL) was added, and the mixture was stirred vigorously for 3 h at room temperature. The layers were separated, and the aqueous portion was washed with ether (5 X 8 mL). The ether extracts were combined with the mother liquors, washed with saturated bicarbonate (5 mL), water (5 mL), and saturated brine (5 mL), dried over anhydrous magnesium sulfate, and filtered. Evaporation of solvent yielded the ketone as a yellow oil (1.2 g, 0.012 mol, 40%), which was purified twice by GC using column a (1 10 "C) and finally column b (70 "C) to yield material which was >99% pure. The GC retention times on both columns matched those of lb: I R (CCI,) 3000 (m), 2910 (m, br), 2310 (w), 2220 (w). 1715 (s), 1430 (m), 1360 (s), 1160 (s) cm-'; IH N M R (CCI,) 6 2.02 (3 H , s), 2.15-2.60 (4 H, m), 5.70 (1 H, br s); 2H N M R (pentane) 6 5.0 (s); MS, M / e (relative intensity) 43 (loo), 57 (lo), 40 (7), 100 (7). 5-Metbyl-5-hexen-2-one ( l a ) was prepared by condensation of ethyl followed by ester hydrolysis acetoacetate with 3-chlor0-2-methylpropene~~ and decarboxylation with refluxing 5% sodium hydroxide. The product was purified by two distillations through a 15411. Vigreux column (bp 142-145 "C) followed by one distillation through a 12-in. spinning band column. The fraction boiling at 142-143 "C was >99% pure by GC. 1-Octene (99.5%) was used as obtained from Albany Chemicals Co. Apparatus. All photolyses were conducted on a preparative scale in pentane (Baker Photrex-grade) solution in a Type K reactor. It consists of a quartz immersion well (55 X 100 cm) fitted with a Suprasil (38 X 98 cm) inner sleeve. The light source was a 40-W American Ultraviolet Co. G3716VH low-pressure mercury resonance lamp, which was encased by a Vycor tube (T254= 70%) for the 254-nm photolyses. The lamp was cooled with a stream of nitrogen, and the entire apparatus was immersed in a large Pyrex container through which cold water (ca. 15 "C) was flowed. Procedure. A 0.05 M solution of the ketone in pentane was placed in the irradiation vessel and was flushed with dry nitrogen for 30-45 min. Methylcyclohexane or isooctane (Baker Photrex grade, 0.2 equiv) was included as internal standard in quantitative runs. Photolyses were generally carried to conversions of